Become a founding member.

← All endpoints
API endpoint

Captioned short

9:16 short from one prompt, slow zoom, one fixed caption

A free diver descends a guide rope into a blue sinkhole under shafts of sunlight, with a burned-in caption, slow zoom.
What goes in
  • clip_promptWhat the picture shows
What it costs
about 5 credits
Made with
GPT Image 2.5, Video Editor

The call

The call behind the sample above, with its real inputs, then the wait for the file. Set WIREFLOW_API_KEY first. Plain HTTP, no SDK to install.

curl -X POST https://www.wireflow.ai/api/v1/apps/captioned-short/execute \
  -H "Authorization: Bearer $WIREFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": {
    "clip_prompt": "Vertical 9:16 photograph from deep underwater looking up: a lone free diver in a black wetsuit descending along a white guide rope into a blue sinkhole, shafts of sunlight cutting through the water from the surface far above, tiny silver bubbles, the walls of the hole fading into dark, shot on film with fine grain"
  }
}'

# It returns an executionId right away. Poll until it finishes.
curl https://www.wireflow.ai/api/v1/apps/captioned-short/executions/EXECUTION_ID/poll \
  -H "Authorization: Bearer $WIREFLOW_API_KEY"

The key needs Full access or Run published apps (apps:execute).

Change this

Same endpoint, different values. Each one starts a run, and you wait for it the same way as the call above.

Change the scene

Say what is in the frame and where the light comes from.

curl -X POST https://www.wireflow.ai/api/v1/apps/captioned-short/execute \
  -H "Authorization: Bearer $WIREFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": {
    "clip_prompt": "Vertical photograph of a glassblower at a furnace at night, orange light on her face, sparks in the air, shot on film"
  }
}'

Change the style

The prompt sets the look too, not only the subject.

curl -X POST https://www.wireflow.ai/api/v1/apps/captioned-short/execute \
  -H "Authorization: Bearer $WIREFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": {
    "clip_prompt": "Vertical ink and watercolor illustration of a lighthouse in a storm, loose brush strokes, one spot of yellow light"
  }
}'